diff --git a/.memory/worklog.json b/.memory/worklog.json
index 192c7da..ced2950 100644
--- a/.memory/worklog.json
+++ b/.memory/worklog.json
@@ -3296,6 +3296,19 @@
       "type": "session-heartbeat",
       "message": "Claude 会话活跃 · 最近命令：claude · 4 项未提交变更 · 最近提交：auto-save 2026-05-14 04:59 (~3)",
       "files_changed": 4
+    },
+    {
+      "ts": "2026-05-14T05:05:22+08:00",
+      "type": "commit",
+      "message": "auto-save 2026-05-14 05:05 (~6)",
+      "hash": "f2663eb",
+      "files_changed": 6
+    },
+    {
+      "ts": "2026-05-13T21:08:50Z",
+      "type": "session-heartbeat",
+      "message": "Codex 会话活跃 · 最近命令：codex · 3 项未提交变更 · 最近提交：auto-save 2026-05-14 05:05 (~6)",
+      "files_changed": 3
     }
   ]
 }
diff --git a/api/main.py b/api/main.py
index b23559e..680c155 100644
--- a/api/main.py
+++ b/api/main.py
@@ -2762,7 +2762,7 @@ def update_storyboard(job_id: str, idx: int, req: UpdateStoryboardReq) -> Job:
 
 
 class PushStoryboardImageReq(BaseModel):
-    kind: Literal["keyframe", "cutout"]
+    kind: Literal["keyframe", "cutout", "asset"]
     frame_idx: int
     element_id: str | None = None
     cutout_id: str | None = None
diff --git a/docs/source-analysis.html b/docs/source-analysis.html
index f6e98ae..2c77c8b 100644
--- a/docs/source-analysis.html
+++ b/docs/source-analysis.html
@@ -839,6 +839,18 @@ SubjectAsset {
         <h2>变更记录</h2>
         <p>这个记录不是 git log 的替代品。它记录“产品理解发生了什么变化、影响了哪些源码、你以后描述需求时该怎么说”。后续每次改功能都要补一条。</p>
         <div class="changelog">
+          <article class="change">
+            <header>
+              <h3>2026-05-14 · 画面工作台增加场景图和主体资产包</h3>
+              <span class="tag violet">Visual Lab</span>
+              <span class="tag blue">Assets</span>
+            </header>
+            <div class="body">
+              <p><strong>问题：</strong>抽帧阶段已经筛过图，画面工作台第一步应把已选关键帧转成可用于生视频的干净素材：每帧一张场景图，同一主体一套多视角/动作/表情图，而不是继续手动逐张抠普通 cutout。</p>
+              <p><strong>改动：</strong><code>KeyFrame</code> 新增 <code>scene_assets</code> 和 <code>quality_report</code>；<code>KeyElement</code> 新增 <code>subject_kind</code> 与 <code>subject_assets</code>。后端新增 <code>generateSceneAsset</code> 和 <code>generateSubjectAssets</code>，主体资产支持白/黑背景、原尺寸/固定尺寸、物体视角、人物/生物动作与喜怒哀乐等表情；当已选关键帧共同指向一个主体时，前端会把这些帧作为 <code>source_frame_indices</code> 传入，后端拼接参考板。</p>
+              <p><strong>影响：</strong><code>api/main.py</code>、<code>web/lib/api.ts</code>、<code>web/components/lightbox.tsx</code>、<code>web/components/nodes/index.tsx</code>、<code>docs/source-analysis.html</code>。生成出的素材保存在 <code>jobs/&lt;jobId&gt;/assets</code>，可作为 <code>asset</code> 类型复制到后续分镜槽位。</p>
+            </div>
+          </article>
           <article class="change">
             <header>
               <h3>2026-05-14 · 抽帧精度自动选择并支持后端排队</h3>